Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
An introduction
to language and society
FOURTH EDITION
PENGUIN BOOKS
SOCIOLINGUISTICS: AN INTRODUCTION
TO LANGUAGE AND SOCIETY
Advisory Editor: David Crystal
Peter Trudgill
Fourth Edition
PENGUI N BOOKS
PENGUIN BOOKS
Penguin Books Ltd, Registered Offices: 80 Strand, London WC2R ORL, England
www.penguin.com
Figures
l Social and regional dialect variation 30
2 Social and regional accent variation 32
3 Sex differentiation in Darkhat Mongolian 68
4 Social-class and style differentiation of non-prevocalic
/r/ in New York City (after Labov) 88
Maps
l Non-prevocalic /r/ in yard and farm in conservative rural
dialects in England 149
2 Modern English dialect areas 152
3 Modern North American dialect areas 154
4 Uvular r in Europe 160
Tables
l Attitudes towards and use of non-prevocalic /r/: upper
middle class in New York City rr
2 Regional and caste differences in Kannada 26
3 RP and local-accent pronunciation of home 32
4 Verbs without -s in Norwich and Detroit 33
5 Non-RP forms for three consonants in Norwich 37
6 New York vowels in bad 38
7 Self-evaluation of tune in Norwich 76
8 Over- and under-reporting of tune in Norwich 76
9 Over- and under-reporting of ear in Norwich 77
ro -in ' forms used in four contextual styles in Norwich 87
Acknowledgements
c chew
German ich, Scots nicht, RP* huge
<;I retroflext d
6 this
g guy
j you
j just
! retroflex l
t retroflex flap, as in some Indian languages and some types
of Swedish and Norwegian
Q. retroflex n
syllabic nasal
IJ sing
1 RP row
R French rose
s she
0 thing
x German nach, Scots loch, Spanish bajo
z vision
1 a glottal stop, e.g. 'cockney' better 'be' er'
\: pharyngeal fricative, as in Arabic
a French patte, North of England pat, Australian part
a RP path, part
re RP pat
e Scots ate, French et
E RP bed
a about
3 RP bird (Note: no [r])
xii Sodolinguistics
RP eat, French il
RP it
close, central unrounded vowel
o French eau, Scots no
:> RP law
e a central vowel between 121 and o
n RP on
121 French eux, German bose
u RP fool, French ou
u RP pull
t:t a central vowel benyeen [ y] and [u] , cf. Scots 'hoose'
A RP up
y French tu, German Uber
vowel nasalized, e.g. o
+ vowel fronted, e.g. [9]
vowel raised, e.g. [Q]
long vowel, e.g. o:
most people will think that. they have a good idea of what it
means. In fact, though, it is not a particularly easy term to define
- and this also goes for the two other commonly used terms
which we have already mentioned, language and accent.
Let us confine our attention for the moment to the terms
dialect and language. Neither represents a particularly clear-cut
or watertight concept. As far as dialect is concerned, for example,
it is possible, in England, to speak of 'the Norfolk dialect' or 'the
Suffolk dialect'. On the other hand, one can also talk of more
than one 'Norfolk dialect' - 'East Norfolk' or 'South Norfolk',
for instance. Nor is the distinction between 'Norfolk dialect'
and 'Suffolk dialect' so straightforward as one might think. If
you travel from Norfolk into Suffolk, the county immediately
to the south, investigating conservative rural dialects as you
go, you will find, at least at some points, that the linguistic
characteristics of these dialects change gradually from place to
place. There is no clear linguistic break between Norfolk and
Suffolk dialects. It is not possible to state in linguistic terms
where people stop speaking Norfolk dialect and start speaking
Suffolk dialect. There is, that is, a geographical dialect continuum.
If we choose to place the dividing line between the two at the
county boundary, then we are basing our decision on social (in
this case local-government-political) rather than on linguistic
facts.
The same sort of problem arises with the term language. For
example, Dutch and German are known to be two distinct
languages. However, at some places along the Netherlands
Germany frontier the dialects spoken on either side of the border
are extremely similar. If we choose to say that people on one
side of the border speak German and those on the other Dutch,
our choice is again based on social and political rather than
linguistic factors. This point is further emphasized by the fact
that the ability of speakers from either side of the border to
understand each other will often be considerably greater than
that of German speakers from this area to understand speakers
of other German dialects from distant parts of Austria or
4 Sociolinguistics
prevocalic /r/ has been very much on the increase in the city in
the speech of the upper middle class. The impetus for this change
may have come from the influx into the city during the war
of many speakers from areas where non-prevocalic /r/ was a
standard or prestige feature, but the change is more clearly due
to a relatedshift in subjective attitudes towards pronunciations
of this type on the part of all New York City speakers. During
the course of the investigation tests were carried out on the
informants' subjective attitudes in order to see if they reacted
to non-prevocalic /r/ as a prestige feature. Those whose response
indicated that for them /r/ was a prestige marker were labelled
'r-positive'. Table I shows the percentage of upper-middle-class
speakers in three age-groups who were 'r-positive' together with
the average percentage of non-prevocalic /r/ used in normal
speech by the same three groups. It can be seen that for speakers
ged under forty there has been a sharp increase in the favourable
evaluation of non-prevocalic /r/. There has, correspondingly,
been an even sharper increase in the use of this /r/ amongst
younger speakers. Other evidence suggests that the change in
subjective attitudes has been the cause rather than the effect of
the change. The change in subjective attitudes, that is, has led
to a change in speech patterns, although it is in fact only the
upper middle class which has made a significant change in its
speech.
% r-positive %/r/
age informants used
8-19 IOO 48
20-39 IOO 34
40+ 62 9
the fact that the island accent is different, but the awareness did
not extend to recognition of the significance of the diphthong
itself. Unconsciously, however, speakers were aware of the social
significance of this pronunciation, and their attitudes towards
it were favourable because of their social attitudes. In other
words, linguistic change does not always take place in the direc
tion of the prestige norm. On the contrary, all sorts of other
attitudes towards language have to be taken into consideration.
Language can be a very important factor in group identification,
group solidarity and the signalling of difference, and when
a group is under attack from outside, signals of difference may
become more important and are therefore exaggerated.
In the following chapters we shall examine some of the com
plex inter-relationships between language and society, of which
subjective attitudes are just one facet. These inter-relationships
take many forms. In most cases we shall be dealing with the
co-variation of linguistic and social phenomena. In some cases,
however, it makes more sense to consider that the relationship
is in one direction only - the influence of society on language,
or vice versa. We can begin with an example of this one-way
relationship which supposedly involves the effect of language on
society. There is a view, developed in various forms by different
linguists, which is most frequently referred to as the 'Sapir
Whorf hypothesis', after the two American anthropologists and
linguists, Edward Sapir and Benjamin Lee Whorf, with whose
names it is most often associated. The hypothesis is approxi
mately that speakers' native languages set up series of categories
which act as a kind of grid through which they perceive the
world, and which constrain the way in which they categorize
and conceptualize different phenomena. A language can affect
a society by influencing or even controlling the world-view of
its speakers. Most languages of European origin are very similar
in this respect, presumably because of their common genetic
relationship and, more importantly, the long cultural contact
between them and the societies in which they are spoken; the
world-views of their speakers and their societies are perhaps for
14 Sodolinguistics
cami camimita
'it is slashed inwards, 'it is fringed'
from side to side'
hani haririta
'it is bent in a rounded 'it lies in a meandering
angle' line'
paci pacicita
'it is notched' 'it is serrated'
roya toyayata
'it makes a turn 'it is rotating'
the other hand, whereas English employs the term uncle for
father's brother and mother's sister's husband, as well as for mother's
brother and father's sister's husband, Njamal uses mama for the
first pair and another term, kama, for the second. Other Njamal
kinship terms distinguish not generation, as in English, but
generation distance. For example, a man can use the same term,
maili, for his father's father and his daughter's son 's wife's sister,
the point being that the person in question is two generations
removed and that these kinship terms are therefore reciprocal -
These, then, are some of the ways in which society acts upon
language and, possibly; in which language acts upon society.
We have seen that there are a number of ways in which language
and society are inter-related, and in the following chapters we
shall investigate some further aspects of this kind of inter
relationship; In the past forty years or so, increasing recognition
of the importance of this relationship has led to the growth of
a sub-discipline within linguistics: sociolinguistics. It is a broad
but fair generalization to say that much of linguistics before
then completely ignored the relationship between language and
society. In most cases this was for good reasons. Concentration
on the 'idiolect' - the speech of one person at one. time in
one style - was a necessary simplification that led to several
theoretical advances. However, as we have already indicated,
language is very much a social phenomenon. A study of language
t<;>tally without reference to its social context inevitably leads
to the omission of some of the more complex and interesting
aspects of language and to the loss of opportunities for further
theoretical progress. One of the main factors that has led to the
growth of sociolinguistic research has been the recognition of
the importance of the fact that language is a very variable
phenomenon, and that this variability may have as much to do
with society as with language. A language is not a simple, single
code used in the same manner by all people in all situations, and
linguists now understand that it is both possible and beneficial to
try to tackle this complexity.
Sociolinguistics, then, is that part of linguistics which is con
cerned with language as a social and cultural phenomenon. It
investigates the field of language and society and has close
connections with the social sciences, especially social psy
chology, anthropology, human geography, and sociology. The
study of attitudes to forms of language, such as the use of
non-prevocalic /r/, is an example of the sort of work carried out
under the heading of the social psychology of language. The study
of Nj amal kinship terms, on the other hand, is a good example
of anthropological linguistics, while the study of the way in which
22 Sociolinguistics
Brahmin . non-Brahmin
Dharwar Bangalore Dharwar Bangalore
'it is' ada ide ayti ayti
'inside' -olage -alli -aga -aga
infinitive affix -likke -ok -ak ak
participle affix -o -o -a -a
'sit' kiit- kut- kunt- kunt-
reflexive ko ko kont- kont
l
1 ....
._ -- ...
""l.,.._---- regional variation -----11
lowest class :
most localized
non standard
In the top line of this table there is only one variant, while there
are eight on the bottom line. In the second line, moreover,
the presence of [ho:m] in both Edinburgh and Newcastle, and,
particularly, of [houm] in both Liverpool and Bradford, indicates
the way in which certain non-RP pronunciations acquire the
status of less locally restricted, regional standard pronunciations
in various parts of the country.
We have known for a long time about this kind of social and
regional dialect and accent variation, and we have also been
fairly well provided with descriptions of RP. We did not know,
however, until work of this sort started, exactly how RP and the
localized j em
Language and Sodal Class 33
It was found in the Detroit survey that there was a clear relation
ship, again of the relative-frequency type, between employment
of the third possibility - double or, better, multiple negation - and
social class. The percentages of nonstandard forms used were:
UMC 2
LMC II
uwc 38
LWC 70
UC 2.7
MC 2. 5
LC 2.3
MMC 1.9
LMC 2.1
uwc 2.8
MWC 2.9
LWC 3 .0
vowels, in order to stress the fact that they do make the distinc
tion, with the result that higher vowels occur in coffee, dog
[dudg] . These high vowels are not the result of pressures of this
sort, since high vowels are by no means confined to Jewish
speakers, but they may well have been encouraged by this ethnic
group substratum effect.
A similar kind of substratum effect can be found in the English
of Scotland. Most Scots today tend to think of themselves as
simply 'Scottish', but historically speaking they repre sent
descendants of two distinct ethnic groups. To simplify things
somewhat, we can say that Highland Scots, whose ancestors
came originally from Ireland, were Gaels, and spoke the Celtic
language Gaelic (as many of them still do in the West Highlands
and on the islands of the Hebrides), while Lowland Scots, like
English people, were ultimately of Germanic, Anglo-Saxon
descent. Now that English is spoken by nearly everyone in
Scotland, this difference still survives in the type of English one
can hear in different parts of the country. Lowland Scots speak
either a local dialect of Scots (see more on this in Chapter 7) 1 or
Standard Scottish English with a local accent (or something in
between) . Highlanders, on the other hand, speak either Standard
Scottish English (which the group as a whole initially learnt as
a foreign language) or something not too far removed from this
- not nearly so far from it as the Lowland dialects, in any case.
Highlanders do not normally say I dinna ken, for example, as
Lowlanders might, but rather I don 't know. There is often, how
ever, a certain amount of substratum influence from Gaelic in
the English spoken by Highlanders which may identify them as
coming from the Highlands. Native speakers of Gaelic, of course,
will often have a Gaelic accent in English, but, one can detect
lexical and grammatical differences even in the speech of High
landers who have never spoken Gaelic in their lives. Examples
include differences such as the following:
Language and Ethnic Group 51
He usually be around.
Sometime she be fighting.
Sometime when they do it, most of the problems always be
wrong.
She be nice and happy.
They sometimes be incomplete.
At first sight, this use of be appears to be no different from its
occurrence in certain British dialects, where I be, he be etc.
correspond to Standard English I am, he is. There is, however, a
crucial difference between AA VE and all other varieties of Eng
lish. As the adverbs usually and sometimes in the above sentences
show, invariant be is used in AA VE only to indicate 'habitual
aspect' - it is only used to refer to some event that is repeated
and is not continuous. There is therefore a verbal contrast in
AA VE which is not possible in Standard English.
and
You lot must have moved it, do I wouldn't have fell in.
You might perhaps think that the first thing you notice about
someone when you meet them for the first time is their clothes,
or their voice, or their eyes, or their smile. If you did think this,
however, you would be quite wrong. Actually, the first thing
you notice about somebody when you first meet them is what
sex they are. This is so obvious that we do not even think about
it. The division of the human race into male and female is so
fundamental and obvious that we take it for granted. The fact
that the difference is so basic means that it is hardly surprising
that it is also reflected and indicated in all human languages. It
is a semantic universal which is lexicalized in all the languages
of the world in terms of pairs of such as man-woman, boy-girl,
son-daughter and so on.
Interestingly, however, languages do differ considerably in
the extent to which sex differences are lexicalized. In German,
for example, you have to specify whether a friend is male, Freund,
or female, Freundin. In English you do not. Kinship terms also
vary; for example 'cousin' is not marked for sex in English but
is marked in many other languages like French, where you have
to specify whether you are talking about a cousin or a cousine. It
can also be true of occupational descriptions where, for example,
languages may or may not distinguish between ' actor' - 'actress',
'manager' - 'manageress', etc. This issue has been the subject of
much controversy recently, witness discussions in the English
speaking world as to whether a woman may be a chairman or
not, and in the French-speaking world as to whether a female
firefighter should be called a pompiere or not. (We return to this
issue in Chapter 10.)
62 Sociolinguistics
The men have a great many expressions peculiar to them, which the
women understand but never pronounce themselves. On the other
hand the women have words and phrases which the men never use,
or they would be laughed to scorn. Thus it happens that in their
conversations it often seems as if the women had another language
than the men.
The savage natives of Dominica say that the reason for this is that when
the Caribs came to occupy the islands these were inhabited by an
male female
'He is saying' /ka:s/ /ka/
'Don't lift it! ' /lakauCi:s/ /lakaucin/
'He i s peeling it' /mols/ /moll
'You are building a fire' /o:sc/ /o:st/
changes had taken place in the male variety which had not been
followed through (or were only just beginning to be followed)
in the women's speech. The same sort of phenomenon occurs
in other languages. Chukchi, for example, is a language spoken
in eastern Siberia. In some dialects, the female variety has inter
vocalic consonants in some words, particularly /n/ and /t/,
which are not present in male forms; for example, male: /nit
vaqaat/; female: /nitvaqenat/. Loss of intervocalic consonants
is a much more frequent and expected sound change than the
unmotivated insertion of consonants, and very many examples
of loss of consonants in this position have been attested in lan
guages from all parts of the world. This kind of distinction would
therefore appear to provide a clear indication that the female
variety is older than the male dialect. In more than one language,
therefore, women's speech is more conservative than that of men.
Another clue comes, again, from Koasati, and in particular
from the attitudes which the Koasati people themselves had to
the two varieties. Older speakers, particularly the men, tended
to say, when asked, that they thought the women's variety was
better than that used by men. This is important, because it ties
in in an interesting way with data we have from technologically
more advanced speech communities (see below) . It also shows
us that the gender varieties are not simply different: in at least
two languages the male varieties are innovating and the female
conservative, and in one case the female variety is evaluated as
better as opposed to worse. Differences of this type should be
easier to explain than linguistic differences, pure and simple.
Let us now take this discussion a stage further by examining
some sex differences in English, where the differences are gener
ally much smaller, less obvious and more subconscious. There
are, it is true, a number of words and phrases which tend to be
sex-bound. (Most of these, incidentally, seem to be exclamations
of some sort. This suggests that taboo may be involved in some
way: it is certainly traditionally more acceptable in our society
for men to swear and use taboo words than it is for women.)
Mostly, however, differences within English are phonetic and
70 Sodolinguistics
In the case of the LMC and the LWC these differences are
Language and Sex 71
very big indeed: men are much more likely to say I don 'twantnone
than women are. Women, this suggests, are far more sensitive to
the stigmatized nature of this grammatical feature than men.
This sensitivity, moreover, is not confined to grammatical fea
tures. In the speech of Detroit Blacks, for instance, women use
a far higher percentage of non-prevocalic /r/ (a prestige feature
here as in New York) than men, allowing for social class:
Female o 3 68 81 97
The results, giving the percentage of boys and girls using each
variant in each case, are given below:
RP Non-RP
gate [ge1t] [g31t]
boys 0 IOO
girls 62 38
can't [ka:nt] [kn:nt]
boys 0 IOO
girls 62 38
out [aut] [aut] [ut]
boys 2S I7 s8
girls 8s IS 0
boy [b::n] [b:;n] [bm] [bm]
boys 0 I6 42 42
girls IS 38 47 0
The boys, we can see, are much more likely than the girls to
use nonstandard local pronunciations.
In different parts of the English-speaking world, then, as well
as in Koasati, female speakers have been found to use forms
considered to be 'better' or more 'correct' than those used by
men. This finding has also subsequently been replicated for large
numbers of other languages in Europe and elsewhere. (There is,
Language and Sex 73
men and half women. But of the 16 per cent over-reporters, all
were women. The figures for the sample as a whole are given in
Table 8. I
Male informants, we earl see, are strikingly more accurat
than their female counterparts. The women, we can say, report
themselves, in very many cases, as using higher-class variants
than they actually do - presumably because they wish they did
use them or think they ought to and perhaps, therefore, actually
believe that they do. Speakers, that is, report themselves as using
the form at which they are aiming and which has favourable
connotations for them, rather than the form they actually use.
(No conscious deceit is involved, it seems.)
Consider, now, the figures in Table 9. This shows the results
of the self-evaluation test for the vowel in Norwich English in
ear, here, idea. (There are two main variants of this vowel in
Norwich: r. [1a] , as in RP, and 2. [:] , with the vowel of care, so
that ear and air, here and hair are the same.) This table shows
not only that a majority of women reported themselves as using
RP [1a] when in fact they did not, but also that as many as half
the men went the other way and under-reported - they reported
Female o I 68 71 83
Language and Sex 79
age-group %
I0-29 55
30-49 63
5 0-69 67
70+ 93
difficult for people who know English to see in what way the
following pairs of sentences differ:
was being studied, and were therefore 'on their guard' as far as
their pronunciation was concerned. This style of pronunciation
has therefore been termed formal speech. In certain parts of the
interview, however, attempts were made to elicit other styles.
At one point, for example, the formality of style was increased
by asking the informants to read aloud from a specially prepared
reading passage. This produced a style that was even more for
mal, because reading aloud is a special case, as it were, of written
rather than spoken language and, secondly, because reading is
a specialized linguistic activity where speakers pay considerable
attention to the way they are speaking. Then the informant also
read aloud from a list of individual words. Here the pronunci
ation was a degree more formal again, since the attention of the
reader was concentrated on a single word at a time, a much
simpler reading task. In this way, then, three different formal
styles of pronunciation were obtained.
What, however, of 'normal', informal speech? Attempts were
made to elicit, in spite of the artificial interview situation, normal
casual speech such as the informant would use in everyday
conversation with friends and family. Several ways emerged in
which this could be done. Casual speech might occur, in the
first place, outside the context of the interview, as in conver
sation with other members of the family who might be present,
or in breaks for a coffee or beer. And it was also found that
certain questions asked during the interview itself were likely
to produce casual speech as a response. Labov, for example,
asked his informants if they had ever been in a situation where
they thought they were in danger of being killed. Generally
informants who related such an incident became emotionally
involved in the narrative and, in attempting to convince the
interviewer of the reality of the danger, forgot the formal ,con
straints of the interview situation. In this way four different
styles of pronunciation were obtained ranging from the infor
mal, casual speech, through formal speech and reading-passage
style, to the most formal, word-list style. This means that scores
obtained by informants for particular linguistic variables can be
Language and Context 87
Table 10. -in ' forms used fo four contextual styles in Norwich
WLS (%) RPS (%) FS (%) CS(%)
MMC 0 0 3 28
LMC 0 IO 15 42
UWC 5 IS 74 87
MWC 23 44 88 95
LWC 29 66 98 IOO
% /rt pronol.llCed
80
eo LMC
UMC
40
. uwc
MWC
LWC
20
Lower
class
CS FS RPS WLS
style
familiar polite
French tu vous
Italian tu Lei
Spanish tU usted
German du Sie
Dutch jij u
Swedish du ni
Norwegian du De
Greek esi esis
Russian ty vy
It has been argued that, originally, the familiar pronouns were
the normal form of address for single individuals, and the polite
forms either second-person plural or third-person pronouns
(Stage 1 - see p. 92) . However, the habit grew up amongst the
upper classes in medieval times of showing respect for a person
by addressing them with what are now the polite pronouns
(following the French forms, we can refer to the familiar pro
nouns collectively as T, the polite forms as V) . This aristocratic
habit led to a situation where, although the upper classes called
each other V and the lower classes used T amongst themselves,
the upper classes used T to the lower classes who, on the other
hand, called them V (Stage 2) . This can be interpreted as signify-
Language and Context 91
intimate: -na
familiar: -e
plain: -ta
polite: -e yo
deferential: -supnita
authoritative: -so
High Low
Swiss German: Hochdeutsch Schweizerdeutsch
Arabic: colloquial classical
Tamil: colloquial literary
High Low
'I say' [aqu:l] [a?u:l]
'I cannot' [la?astati<t] [ma? dad]
'many' [ka0irah] [katir]
'that' [oa:ka] [da]
English:
'He never actually became Swiss, neither on paper nor in his heart;
and you could tell from his language that he had not grown up there.
It was not only his language that showed that he was a foreigner - his
way of life showed it too. He preferred to associate with his German
compatriots rather than with the natives, and was a member and the
treasurer of their society.'
Standard German:
Als Rodange I872 sein Buch drucken liess, hatte er keinen Erfolg damit.
Mit zuviel List war er ein paar Leu ten aufdie Zehen getreten, und die konnten
ihm das nicht verzeihen. Erst eine Generation spater begann Rodange, seinen
ihm zustehenden Platz zu erhalten. Seine Kinder haben es wenigstens noch
erlebt, dass I927 ein wenigvon demgutgemachtwurde, was an ihm verbrochen
worden war!
English:
'When Rodange had his book printed in 1872 he had no success with
it. With too much intrigue he had trodden on some people's toes, and
they could not forgive him that. Only a generation later did Rodange
begin to receive his rightful place. His children at least experienced the
making good, in 1927, of some of the wrong that had been done him.'
(The two Spanish passages can be translated as: 'that I'm going
to stop smoking because it's harmful to me' and 'and there I go
106 Sociolinguistics
his six-year-old son Todd to carry out the same how come?
questioning routine as above. In his case, relaxed and unembar
rassed exchanges like the following occurred:
Answerer: Hello.
Caller: Is fohn there?
Answerer: Hello.
Caller: Is that r234567?
A: Yes.
C: This is Andre here. I'm sorry to disturb you. Is fean there?
It is normal, that is, for callers to apologize for the intrusion, and
to identify themselves. In American telephone conversations,
n4 Sociolinguistics
the French speak French, and so on. There are good reasons for
this, but the reality of the matter is somewhat different. Nearly
all European countries contain indigenous linguistic minorities
- groups of speakers who have as their native variety a language
other than that which is the official, dominant or major language
in the country where they live. Iceland, where 100 per cent of the
indigenous population are Icelandic-speaking, is the exception
rather than the norm. In some cases, where the minorities are
relatively large, the nation-state usually has more than one
offici;il language. Examples are Belgium (Dutch - often \mown
as Flemish in Belgium - and French), Switzerland (German,
French, Italian and Romansch), and Finland (Finnish and
Swedish) .
Where the minority is smaller or less influential, the minority
language or languages are unlikely to have official status, and
their speakers, often out of sheer practical necessity, will tend
to be bilingual. This last factor is what helps to give Europe its
outwardly monolingual appearance. The overwhelming major
ity of French citizens can speak French, in spite of the fact that
for a number of them it is a second language. The same sort of
situation applies in the United Kingdom. The UK also gives
every appearance of being monolingual, and visitors certainly
need to learn no other language than English. This appearance,
though, is somewhat deceptive. It is true that England has not
had an indigenous linguistic minority since Cornish became
extinct in the eighteenth century, but there are living in the
country today sizeable groups of speakers of very many lan
guages from a number of different places around the world,
including, for example, the northern Indian subcontinent, such
as Punjabi and Bengali (and there are also some grounds for
arguing that the first language of many older British people of
West Indian origin is not English, although it is very similar -
see Chapter 9) . Of the indigenous languages, Welsh is the first
language of about a fifth of the population of Wales, while Scots
Gaelic is spoken natively by about 70,000 people, largely in the
West Highlands and Hebridean Islands of Scotland.
Language and Nation 121
Kurds do not exist'); and it can very seriously impair the edu
cational progress of children who have to learn a new language
before they can understand what the teacher is saying, let alone
read and write.
This approach was also for many years the policy in the
United States, where it may have been at least partly responsible,
together with the broader social attitudes to minority languages
that went with it, for the widespread and rapid assimilation of
minority language groups to the English-speaking majority. (For
the more recent so-called 'English Only' movement in the USA,
see Chapter ro.) Today, considerable provision is made for some
minority groups, notably Spanish-speakers and Native American
Indians, to be educated in their own language, and certain other
steps have also been taken: public notices in New York City, for
example, are posted in Spanish as well as English, to cater for
the large Puerto Rican community now living there. However,
even the larger, more rural linguistic minorities such as those
consisting of speakers of French (in the North-East and in Louisi
ana) and Pennsylvania Dutch (a form of German) are rapidly
declining in size. In 1970, the ten largest linguistic minorities in
the USA were as follows:
Spanish 7.9 million Yiddish 1. 5 million
German 6.2 Norwegian o.6
Spanish:
Maigret escucha distraidemente, pensando que medio Paris esta de vaca
ciones y que el resto a estas horas estara tomando refrescos en las mesitas de
las terrazas. Que condesa? Ah, si! El hombre triste s'explica. Una seflora que
ha sufrido mas de un contratiempo y que abri6 un salon de bridge en la
calle Piramides. Una mujer muy guapa. Se conoce que el pobre hombre esta
enamorado. - Hoy, a las cuatro, he cogido un billete de mil de la caja de los
duefios.
English:
'Maigret only half listens, thinking that half Paris will be on holiday
and that the rest, at this hour, will be drinking cool drinks outside at
small cafe-tables on the pavement. Which countess? Ah yes! The sad
man explains. A lady who has had more than one setback and who has
opened a bridge-dub on Pyramides Street. A rather pretty woman. It is
apparent that the poor man is in love with her. "Today, at four o'clock,
I took a thousand-franc note from the bosses' till. " '
this was still the language o.f the influential urban elite. In
fact, the position of Dano-Norwegian was strengthened when
teachers were instructed, in 1887, not to teach the reading pro
nunciation of Danish but rather the colloquial standard, the
modified Dano-Norwegian Riksmal. These two acts were the
government's first involvement in language planning.
The origin of the conservative and radical forms in the two
official languages today lies in the desire of successive govern
ments to establish one national language instead of two without
actually abolishing either of them. Rather the desire has been
to reform the two gradually towards each other. For example,
Norwegian dialects, including those of the urban working-class,
have three genders for nouns (masculine, feminine and neuter),
whereas Dano-Norwegian, subsequently Riksmal, had, like
Danish, only two (common and neuter) . This meant that
Rksmal had identical forms for the definite article (which in
Norwegian is placed after the noun) for masculine and feminine
words: mann 'man', ko 'cow'; mannen 'the man', koen 'the cow'.
Landsmal had distinct forms: kui 'the cow' . In 1917 the govern
ment introduced an official reform, one of the effects of which
was to achieve a compromise between the two languages on
this (and other) points. In Landsmal the form of the feminine
definite article was to be changed from -i to -a to bring it into
line with eastern dialects, while in Riksmal the feminine form
-a was introduced obligatorily for some words, particularly words
with rural associations, like cow, and optionally for others. This
meant that 'the cow' was now kua in both languages. ('Obliga
tory' here means obligatory in school textbooks and in school
children's writing.) As a result of this reform the feminine
definite article used in conjunction with some nouns in Bokmal
is considered to be a radical form, the masculine (or common)
article a conservative form.
The next important development in government language
planning was the 1938 reform, which was based on the report
of a committee whose mandate was 'to bring the two languages
closer together with respect to spelling, word-forms, and
140 Sodolinguistics
Conservative Nynorsk: Da t
Moderate Nynorsk: Da
Radical Nynorsk: Da ho vakna morgonen etter, kjendest
-
Radical Bokmal: Da ho vakna morgenen etter, f0ltes
Moderate Bokmal: hun vaknet (vaknet) etter,
Conservative Bokmal: hun vaknet etter, i::
:t
(literal translation) When she awoke the morning after, felt !:l
Cons. Nn. : or augo \"Ii
Mod. Nn. : av auga Sl
=
Rad. Nn. : av auga hans. Sl.
It's noo apen fur tae pit in jab foarms fur tha ontak o Unner-Editor (Inglis
an Ulster-Scotch) wi tha Chaummer o tha Scrievit Account o tha New Ulster
Semmlie sittin at tha Tolsel Biggins, Stormont, Bilfawst. A start wull be gien
fur sax month, wi anither contraick aiblins forbye.
in England
150 Sodolinguistics
General Canadian
....... __ _
Sodolinguistics
UNITED STATES
The West
speakers did not arrive from the British Isles in large numbers
until the 1840s. Fortunately we have some recordings of what
people born in New Zealand in the period from 1850 to 1890
sounded like - large numbers of them were recorded in the 1940s
for radio broadcasts of pioneer reminiscences. What we can
learn from this is the following. The first generation of New
Zealand-born anglophones grew up in situations where a whole
mixture of dialects from different parts of the mother country
were spoken, and they seemed to have pleased themselves (with
out of course giving it any conscious thought) about which
_
vowels and consonants they would select from the mixture
present in the adult speech around them. They therefore used
new combinations of features - one woman pronounced here
and there as if she were Scottish, and out as if she came from
London! - and these combinations would differ from one person
to another even if they had grown up alongside one another.
The second generation, however, were the ones who carried out
the levelling process: they seem (again totally subconsciously)
to have agreed on a unified, common accent by selecting the
features which were most common in the original mixture.
Modem New Zealand English today therefore does not have the
north of England vowel in words like tough because this was
absent from the English not only of the south of England but
also that of Scotland and Ireland. On the other hand, modem
New Zealanders do not 'drop' the /h/ of words like house because,
although this is normal in the south of England, it was not
found in Eastern England or Scotland or Ireland. We can also
suggest that New Zealand English is very similar to Australian
English because they are both the result of the koineization
of a similar mixture of similar British Isles dialects in similar
proportions.
We saw above that distance is clearly an important factor in
the spread of linguistic forms, although in many cases social
distance may be as important as geographical distance, as we
have noted: two towns may be socially 'closer' to each other
than they are to the intervening stretches of countryside. But
Language and Geography 159
(j
Areas where
uwlar r is
NOT usual in
educated speech
---- Language
boundaries
not coinciding
with national
frontiers
living in the area. This last variety has therefore had a history
something like this:
English
! (pidginization)
West African Pidgin
! (creolization)
Djuka
! (pidginization)
Pidgin Djuka
Both Sranan and Djuka are uncontroversial, socially and lin
guistically. They are recognized by all to be creoles, and as
languages distinct from English: it would be difficult to make
out a good case for the above specimen of Sranan as a type of
English. Mutual intelligibility between Sranan and English is
nil. Socially, too, there are no reasons for regarding Sranan as a
form of English. Dutch is the official language in Sranan, and
English itself is little used. In other parts of the world, however,
the position is much less clear. In parts of West Africa, for
instance, Pidgin English is widely employed as a lingua franca,
and in certain areas, notably in parts of Nigeria, it has become
creolized. The difficulty there is that, in contrast to Surinam,
English is an official language and is used, as a 'world language'
of high prestige, for many different functions throughout the
country. The result of this is that Nigerian Pidgin, even in its
creolized form, has become heteronomous (see p. 4) with
respect to Standard British and/or Nigerian English. Pidgin Eng
lish is subject to considerable influence from English, and is
widely considered simply to be a 'bad' or 'corrupt' form of
English.
In Sierra Leone the situation is similar, although if anything
rather more complicated. In Freetown, the capital, it is possible,
probably somewhat artificially, to distinguish between four dif
ferent linguistic varieties which have some connection with
English:
Language and Contact 1 73
1. British English;
2. Sierra Leone English - spoken mainly by middle-class Sierra
Leonians, and containing certain features due to the influ
ence of African languages;
3. West African Pidgin English - used as a (mainly commer-
cial) lingua franca;
4. Krio.
Krio is an English Creole with about 30,000 native speakers
living in and around Freetown. The language developed from
an English Creole spoken by slaves returned from Jamaica, North
America and Britain, and is not directly connected with West
African Pidgin. The following four versions of the same sentence
illustrate some of the similarities and differences involved:
British English: /mm gomIJ ta W3:k/
Sierra Leone English: /aim goin to wJk/
West African Pidgin: /o di go wJk/
Krio: /o di go wok/
The similarities between the four varieties have led to the con
clusion on the part of many Sierra Leonians that the three lower
prestige forms represent unsuccessful attempts to imitate the
higher prestige British English - and Pidgin and Krio in particular
are often simply regarded as 'broken English'.
In parts of the former British West Indies the position is again
similar, but the problems it brings with it are considerably more
severe. Let us consider Jamaica. Some linguists writing about
the language spoken in Jamaica refer to it as Jamaican English
while others, preferring to give it the status of a separate lan
guage, call it Jamaican Creole. This disagreement about terminol
ogy is the result of the discreteness and continuity problem we
mentioned in Chapter 1 (p. 5) . In Jamaica, Standard English is
the official language and is spoken there, at the top of the social
scale, by educated Jamaicans and people of British origin. At the
other end of the social scale, particularly in the case of peasants
in isolated rural areas, the language used is an English-based
1 74 Sodolinguistics
more or less desirable in any way), then this gives us the following
chronological picture:
Language and Contact 181
Process: source
pidginization Pre-pidgin simplification admixture reduction
focusing Pidgin
creolization Creole expansion
partial Post-creole complication purification
decreolizatiori
further Vestigial complication purification
decreolization Post-creole
its verb phrases are Cree, including the complex verbal mor
phology of that language. For example:
la fam micimine:w li pci
the (fem.) woman she-is-holding-it the (masc.) little-one
French Cree French
'The woman is holding the child.'
10 Language and Humanity
society lays down different roles for men and women, but it is
equally true that what society lays down can and does change
- and will change if enough members of the society feel that it
is desirable that this should happen. In most Western societies,
for instance, many people have altered or are altering the way
they feel about what is appropriate as far as gender roles are
concerned. And these beginnings ofa move away from gender
role stereotyping probably explain the fact that linguistic differ
ences between younger men and women - a very interesting
finding from sociolinguistic research - are statistically smaller
wear can be found for sale alongside men 's wear and so on.
-
have just cited has any basis in fact, but they can have all sorts
of unfortunate consequences.
One of the very distressing consequences that attitudes of this
type can have is language death. One of the questions linguists are
often asked is: how many languages are there in the world? This
is a rather difficult question to answer, not least because of the
dialect-versus-language issue we have discussed a number of
times in this book. It is not too inaccurate to say, however, that
there are about 6,ooo languages in the world today. What is
much more certain is that this number is smaller than it used
to\ be and is getting smaller all the time. In the last years of
the:twentieth century, languages are dying out without being
replaced at an increasingly catastrophic rate.
What happens is that communities go through a process of
language shifi:. This means that a particular community gradually
aandons its original native language and goes over to speaking
another one instead. This has been a relatively common process
in the sociolinguistic history of the world. Two hundred years
ago, for example, most of the population of Ireland were native
speakers of lrish Gaelic. Now the vast majority are native speakers
of English. Before the Roman conquest, the population of much
of what is now France were speakers of the Celtic language
Gaulish. Subsequently, however, they shifted to the language
of their conquerors, Latin, which eventually became French.
Later on, the northern part of France was conquered by the
Germanic-speaking Franks. These conquerors, however, eventu
ally went through a process of language shift and ended up
speaking French too. Similarly, the Norwegian-speaking Vikings
who subsequently conquered and settled in the part of northern
France we now call Normandy also shifted from their Scandinav
ian language to French. A few generations later, as a result of
the Norman conquest of England in rn66, these former Scandina
vians took the French language to England. Once in England,
however, it took the descendants of the Norman conquerors
only a few generations before they shifted once again, this time
to English.
192 Sodolinguistics
made to feel that their native vernacular dialects are inferior, not
only socially, which is unfortunately true, but also linguistically,
which is most emphatically not true. It is hardly surprising,
therefore, if many of them try to shift to the standard variety
even if, at some level of consciousness, they do not really want
to.
In this kind of atmosphere, traditional dialects or patois can
disappear surprisingly quickly. Traditional dialects have more
or less disappeared from most of England, for example - although
not from Scotland - and in many parts of the French-speaking
world the picture is the same. There is often, of course, a direct
relationship between the degree of hostility to dialects and the
rate at which they disappear. One way of combating this hostility
is to point to those fortunate, more tolerant societies which do
have greater respect for language varieties as good examples to
be followed.
In many dialect-hostile parts of Europe, including England,
there is a widespread view that dialects are out-of-date, old
fashioned, unsophisticated, divisive and economically disad
vantageous. To combat this belief, we can point to the following
fact. In 19901 according to many measurements of per capita
income, the three richest countries in Europe were Luxembourg,
Norway and Switzerland; all three countries are dialect-speaking.
As we saw in Chapter 51 the entire indigenous population of
Luxembourg is dialect-speaking. They learn and use German and
French, but their mother-tongue is Luxemburgish/Letzeburgish,
which is widely regarded as a dialect of German. Norway, too,
is one of the most dialect-speaking countries in Europe. Some
people do speak a form of Standard Norwegian, but the majority
do not, whatever the social situation. People speak dialect on
radio and TV, professors give lectures in dialect, and authors
write poems and novels in dialect. The most important aspect
of the Norwegian language situation, however, is, as we saw
in Chapter 71 that there is an enormous societal tolerance fm
linguistic diversity and that, what is more, linguistic diversity
in Norway is officially recognized and officially protected. This
198 Sociolinguistics
Chapter r
A lexicon of sociolinguistic terms is P. Trudgill, Introducing Lan
guage and Society (Penguin) . Useful introductory works to socio
linguistics include W. Downes, Language and Society (Cambridge
UP); J. Holmes, An Introduction to Sociolinguistics (Longman); R.
Fasold, The Sociolinguistics ofLanguage (Blackwell); R. Fasold, The
Sociolinguistics of Society (Blackwell); R. Hudson, Sociolinguistics
(Cambridge UP). There is an emphasis on anthropological lin
guistics in D. Hymes (ed.), Language in Culture and Society (Harper
& Row), which contains a paper by Haas on interlingual taboo
from which I have taken some of the data used here; and in R.
Burling, Man's Many Voices (Holt, Rinehart & Winston), where
the full Njamal kinship data occurs; in W. Foley, Anthropological
Linguistics (Blackwell); and in A. Duranti, Linguistic Anthropology
(Cambridge UP). The writings of Sapir and Whorf in this area
are to be found in D. Mandelbaum (ed.), Selected Writings of
Edward Sapir in Language, Culture and Personality (California UP),
and B. Wharf, Language, Thought and Reality (M IT Press) . Work
on geolinguistics is discussed in J. K. Chambers and P. Trudgill,
Dialectology (Cambridge UP), while secular linguistics is the
topic ofJ. Chambers, Sociolinguistic Theory (Blackwell) and of W.
Labov, Sociolinguistic Patterns (Pennsylvania UP). The influence
of Labov's work can be noted in much of the present book, and
his writings are both stimulating and important. Also recom
mended are his The Social Stratification of English in New York
City (Center for Applied Linguistics = CAL) and Principles of
Linguistic Change (Blackwell) . Friedrich's study of Russian
206 Sodolinguistics
Chapter 2
The results of th e Detroit urban dialect survey led by Roger Shuy
are not readily available in their entirety. Some of the results,
however, including many of those I have used here, are set out
in W. Wolfram, A Sociolinguistic Description ofDetroit Negro Speech
(CAL) . The Norwich data can be found in P. Trudgill, The Social
Differentiation of English in . Norwich (Cambridge UP). socio
linguistic methodology is discussed in L. Milroy, Observing and
Analysing Natural Language (Blackwell), and in J. Cheshire, Vari
ation in an English Dialect (Cambridge UP).
Chapter 3
The topic of language and ethnicity is the subject of H. Giles
(ed.), Language, Ethnicity and Intergroup Relations (Academic),
and of J. Edwards, Language, Society and Identity (Blackwell) . W.
Whiteley (ed.), Language Use and Social Change (Oxford UP)
contains a paper by Berry dealing with the Accra data, together
with a number of other articles relevant to topics discussed in
Chapters s and 6. Readers interested in AAVE are referred to W.
Labov, Language in the Inner City (Pennsylvania UP), W. Wolfram
and N. Clarke (eds), Black- White Speech Relationships (CAL);
the more popular J. Dillard, Black English (Random House); S.
Mufwene et al., African-American English (Routledge); S. Poplack,
The English History of African American English (Blackwell); and
J. Rickford, African American English (Blackwell) .
Chapter 4
Jespersen's writings on this topic can be found in Language:
Its Nature, Development and Origin (Allen & Unwin), while the
Koasati data is taken from a paper by Haas that appears in the
Annotated Bibliography and Further Reading .207
Chapter s
Brown and Gilman's T- and V-pronoun article, 'The Pronouns
o Power and Solidarity', and Ferguson's 'Diglossia' paper have
both been reprinted a number of times. The former appears in
J. Fishman (ed.), Readings in the Sociology ofLanguage (Mouton),
and the latter in the Hymes reader (see Chapter r notes) . A
number of articles in J. Gumperz and D. Hymes (eds), Directions
in Sociolinguistics (Blackwell) deal with aspects of the relationship
between language and social context. Also relevant are J. Fish
man, Language in Sociocultural Change (Stanford UP), S. Romaine,
Bilingualism (Blackwell) and J. Gumperz, Language in Social
Groups (Stanford UP). The relationship between style, register
and dialect is discussed in L. Andersson and P. Trudgill, Bad
Language (Penguin) .
Chapter 6
Aspects of the social psychology of language, including the study
of the use of language in social interaction and the findings of
matched guise experiments, can be found in H. Giles and R.
St Clair (eds), Language and Social Psychology (Blackwell) . The
study of conversation is dealt with in M. Coulthard, Introduction
to Discourse Analysis (Longman), and the more advanced M.
208 Sodolinguistics
Chapter 7
The discussion of language planning in Malaysia is taken from
R. Le Page, The National Language Question (Oxford UP). Many
of the facts about the Norwegian situation can be found in a
detailed treatment by E. Haugen, Language Conflict and Language
Planning: The Case of Modem Norwegian. Good sources of infor
mation on problems of planning and standardization in differ
ent parts of the world include J. Fishman et al. (eds), Language
Problems o(Developing Nations (Wiley), R. Cooper, Language Plan
ning and Social Change (Cambridge UP), and }. Rubin et al. (eds),
Language Planning Processes (Mouton) . For information on the
linguistic situation in the English-speaking world, see C. Fer
guson and S. B. Heath (eds), Language in the USA (Cambridge
UP) and P. Trudgill (ed.), Language in the British Isles (Cambridge
UP).
Chapter 8
The English dialect data presented can be found in the publi
cations of the Survey of English Dialects edited by H. Orton et al.
and published by E. J. Arnold. The basic material is published
in four volumes, each in three parts (i.e. twelve books in all) : I.
The Six Northern Counties; I I . The West Midland Counties;
A nnotated Bibliography and Further Reading 209
I I I . The East Midland Counties and East Anglia; and IV. The
Southern Counties. (These are best consulted in libraries.) More
accessible information is available in M. Wakelin, English Dia
lects (Athlone), P. Trudgill, Dialects (Routledge), A. Hughes and
P. Trudgill, English Accents and Dialects (E. J. Arnold); and K. M.
Petyt, The Study ofDialect (Deutsch) . Map 2 is based on P. Trudgill,
The Dialects ofEngland (Blackwell) . Dialect mixture and koineiz
ation is discussed in P. Trudgill and D. Britain, Dialects in Contact
(J3lackwell) .
Chapter 9
The Tok Pisin data is taken from R. Hall, Pidgin and Creole
Languages (Cornell UP). L. Todd, Pidgins and Creoles (Routledge
& Kegan Paul) is an introductory text, as are L. Todd, Modern
Eglishes (Blackwell), and S. Romaine, Pidgin and Creole Languages
(Blackwell) . Bailey's data can be found in an article in D. Hymes
(ed.), Pidginization and Creolization ofLanguages (Cambridge UP).
My 'Dutch-Swedish Pidgin English' example is based on a paper
by Whinnom in the same volume. The subject of pidgins and
creoles has become increasingly popular and well researched in
recent years, and many valuable publications are now available.
Among the most interesting is I. Hancock, Readings in Creole
Studies (Story-Scientia) . Bickerton's important book is Roots of
Language (Karoma) .
Chapter rn
Corbett's book is G. Corbett, Gender (Cambridge UP). A number
of irrational attitudes about language are discussed in L. Bauer
and P. Trudgill (eds), Language Myths (Penguin) . Dialect in edu
cation is the subject of P. Trudgill, Accent, Dialect and the School
(E. ] . Arnold), and W. Wolfram and D. Christian, Dialects and
Education (Prentice Hall) . J. Fishman, Reversing Language Shift.
(Multilingual Matters) is the seminal text in this field. Language
death is the subj ect of N. Dorian, Investigating Obsolescence:
210 Sociolinguistics
\HJ O NLINE
'
www. pe ng u i n . co m
READ MORE IN PENGUIN
I n every corner o f the world, o n every subject under the sun, Penguin
represents quality and variety - the very best iri publishing today.
In the United Kingdom: Please write to Dept. EP. Penguin Books L td,
Bath Road, Harmondsworth, West Drayton, Middlesex UB7 ODA
In New Zealand: Please write to Penguin Books (NZ) Ltd, Private Bag
102902, North Shore Mail Centre, Auckland 1 0
In South Africa : Please write to Penguin Books South Africa ( Pty) Ltd,
P. 0. Box 751093, Gardenview, 204 7 Johannesburg
READ MORE IN PEN GUIN
'A deliciously filthy trawl among taboo words across the ages and the
globe' Observer. 'Erudite and entertaining' Penelope Lively, Daily
Telegraph